基于卫星数据的强对流初生预警和评价技术研究
|
摘要
强对流灾害天气是我国的主要灾害性天气之一,常对社会的生产、生活构成严重的威胁。近年来,针对强对流初生(Convective Initiation,CI)的监测和预警成为新的热点研究问题。此处,CI的定义为:多普勒天气雷达第一次检测到由对流云产生的反射率因子大于等于35dBZ。可以认为,CI的发生是强对流天气活动开始的标志,因而,对CI进行成功的预警,也就成为了灾害性天气监测预警的重要环节。本文主要包括以下两部分。
第一部分是CI预警算法。具体步骤为:(1)利用CL多小波融合方法将卫星红外数据插值成与可见光等同的分辨率,增加红外通道数据的物理信息量;(2)结合交叉相关算法,追踪15分钟间隔的多幅卫星图像中同一像素的运动趋势,用于估计云顶亮温的时间变化趋势;(3)采用包括了IR云顶亮温,IR多通道差和IR云顶亮温/多光谱时间变化趋势的八个独立的CI判别指标来预警CI,具体为若积雨云像素满足八个CI判别指标中的七个或七个以上,则被标识为具有高度的CI可能性,也就是本文采用的计分统计方法。结果表明通过对对流云关键的IR亮温值和时间变化趋势的监测,可以提前30~45分钟预警出CI。
第二部分是预警算法的评价技术。在这一部分我们用四个评价技术指标—预警成功率(POD)、虚假警报率(FAR)、风险得分(TS)、Heidke技巧得分(HSS),和主成分分析(PCA)方法对本文应用的预警CI算法进行评价,从统计分析的角度评价本文使用的八个指标计分统计方法预警CI的准确性,以及确定每一个指标对于预警CI的相对重要性。通过对统计结果的分析,验证了本文所用的预警CI算法的可靠性。
Convective weather is one of the severe weathers in China and often poses serious threats on the production and life of human beings. Recently, it became a hot topic on monitoring and forecasting convective initiation(CI). CI is defined as the first occurrence of a≥35dBZ radar echo from a cumuliform cloud. CI is the beginning symbol of the strong convection weather event. Thus, the accurate forecasting of CI is important for convective weather warning. This study mainly contains two parts.
The first is the CI forecasting algorithm, which is as follows: (1)Fusing visible imagery with high spatial resolution and rich texture detail to infrared imagery, based on the theory of CL multiple wavelet fusion, which can improve the spatial resolution of infrared imagery, and increase the physical information. (2)Tracking the moving trend of the same pixel in different satellite imagery of 15-min time period by the cross correlation algorithm. The goal is to assess the temporal trend in cloud-top temperature. (3)Eight predictors are used to forecast CI which include IR cloud-top brightness temperatures, IR multispectral channel differences, and IR cloud-top temperature/multispectral temporal trend. Cumulus cloud pixels for which≥7 of the 8 CI indicators are satisfied are labeled as having high CI potential. The result shows that CI may be forecasted 30-45 min in advance through this method.
The second is a statistical evaluation of forecasting CI. Four measures of foresting skill which are probability of detection(POD), false-alarm ratio(FAR), threat score(TS), and Heidke skill score(HSS), and principal component analysis(PCA) are used to evaluate the CI forecast products. The goal of the statistical analysis is to confirm the accuracy of the algorithm, and that whether each CI indicator is important to forecast CI. The results show that it is reliable to use this method.
第一部分是CI预警算法。具体步骤为:(1)利用CL多小波融合方法将卫星红外数据插值成与可见光等同的分辨率,增加红外通道数据的物理信息量;(2)结合交叉相关算法,追踪15分钟间隔的多幅卫星图像中同一像素的运动趋势,用于估计云顶亮温的时间变化趋势;(3)采用包括了IR云顶亮温,IR多通道差和IR云顶亮温/多光谱时间变化趋势的八个独立的CI判别指标来预警CI,具体为若积雨云像素满足八个CI判别指标中的七个或七个以上,则被标识为具有高度的CI可能性,也就是本文采用的计分统计方法。结果表明通过对对流云关键的IR亮温值和时间变化趋势的监测,可以提前30~45分钟预警出CI。
第二部分是预警算法的评价技术。在这一部分我们用四个评价技术指标—预警成功率(POD)、虚假警报率(FAR)、风险得分(TS)、Heidke技巧得分(HSS),和主成分分析(PCA)方法对本文应用的预警CI算法进行评价,从统计分析的角度评价本文使用的八个指标计分统计方法预警CI的准确性,以及确定每一个指标对于预警CI的相对重要性。通过对统计结果的分析,验证了本文所用的预警CI算法的可靠性。
Convective weather is one of the severe weathers in China and often poses serious threats on the production and life of human beings. Recently, it became a hot topic on monitoring and forecasting convective initiation(CI). CI is defined as the first occurrence of a≥35dBZ radar echo from a cumuliform cloud. CI is the beginning symbol of the strong convection weather event. Thus, the accurate forecasting of CI is important for convective weather warning. This study mainly contains two parts.
The first is the CI forecasting algorithm, which is as follows: (1)Fusing visible imagery with high spatial resolution and rich texture detail to infrared imagery, based on the theory of CL multiple wavelet fusion, which can improve the spatial resolution of infrared imagery, and increase the physical information. (2)Tracking the moving trend of the same pixel in different satellite imagery of 15-min time period by the cross correlation algorithm. The goal is to assess the temporal trend in cloud-top temperature. (3)Eight predictors are used to forecast CI which include IR cloud-top brightness temperatures, IR multispectral channel differences, and IR cloud-top temperature/multispectral temporal trend. Cumulus cloud pixels for which≥7 of the 8 CI indicators are satisfied are labeled as having high CI potential. The result shows that CI may be forecasted 30-45 min in advance through this method.
The second is a statistical evaluation of forecasting CI. Four measures of foresting skill which are probability of detection(POD), false-alarm ratio(FAR), threat score(TS), and Heidke skill score(HSS), and principal component analysis(PCA) are used to evaluate the CI forecast products. The goal of the statistical analysis is to confirm the accuracy of the algorithm, and that whether each CI indicator is important to forecast CI. The results show that it is reliable to use this method.
引文
[1] Mecikalski J R,Bedka K M.Forecasting convective initiation by monitoring the evolution of moving cumulus in daytime GOES imagery.Mon. Wea. Rev.,2006,134:49-78
[2] Dixon M , Wiener G . TITAN: Thunderstorm identification, tracking, analysis, and nowcasting a radar-based methodology.Journal of Atmospheric and Oceanic Technology,1993,10(6):785-797
[3] Johnson J T,Mackeen P L,Witt A, et al.The Storm Cell Identification and Tracking algorithm: an enhanced WSR-88D algorithm.Weather and Forecasting,1998,13:263-276
[4] Jan Handwerker.Cell tracking with TRACE3D—a new algorithm.Atmospheric Research 2002,61:15-34
[5] K A布朗宁.现时预报(周凤仙,马振,李泽椿译).北京:气象出版社,1986,260
[6]许锐.基于卫星数据的对流初生自动识别研究:[硕士学位论文].青岛:中国海洋大学,2009
[7] Maddox R A.Mesoscale convective complexes.Bull Amer Meteor Soc,1980,61(11):1374-1387
[8] Setvák M,Doswell C A.The AVHRR channel 3 cloud top reflectivity of convective storms.Mon. Wea. Rev.,1991,119:841-847
[9] Strabala K I,Ackerman S A,Menzel W P,Cloud properties inferred from 8–12 _m data.J. Appl. Meteor.,1994,33:212-229
[10] Schmetz J,S A Tjemkes,M Gube,et al.Monitoring deep convection and convective overshooting with METEOSAT.Adv. Space Res.,1997,19:433-441
[11] Roberts R D,S Rutledge.Nowcasting storm Initiation and growth using GOES-8 and WSR-88D data.Wea. Forecasting,2003,18:562-584
[12] Hand W H . An object-oriented technique for nowcasting heavy showers and thunderstorms.Meteor. Appl.,1996,3:31-41
[13] Nair U S,R C Weger,K S Kuo,R M Welch.Clustering, randomness, and regularity in cloud fields 5. The nature of regular cumulus cloud fields.J. Geophys. Res.,1998,103(D10):11363-11380
[14] Purdom J F W,Menzel W P,Introducing GOES-I: The first of a new generation of geostationary operational environmental satellites.Bull. Amer. Meteor. Soc.,1994,75:757-781
[15] Beckman S K.Relationship between cloud bands in satellite imagery and severe weather. Satellite Imagery Interpretation for Forecasters.1986,Vol. 2,Precipitation Convection, P. S. Parke, Ed.,National Weather Association.[Available from the NWA, 4400 Stamp Road, No. 404, Temple Hills, MD 20748.]
[16]倪允琪,周秀骥,王德英等.我国重大天气灾害形成机理与预测理论研究进展.中国气象科学研究院年报,2002,1:8-12
[17]徐建平.国内外气象卫星发展.空间科学学报,2000,10(20):104-115
[18]陈渭民.卫星气象学(第二版).北京:气象出版社,2005,535
[19]李伟,张恒.气象卫星的应用与发展.卫星与网络,2007,8:56-60
[20]王敏.MTSAT卫星云图中对流初生的自动预报算法研究:[硕士学位论文].青岛:中国海洋大学,2010
[21]石秀云.卫星云图分析基础及应用.青海科技,2010,3:43-46
[22]徐建平.日本MTSAT多功能卫星.气象科技,2005,33(1):96-97
[23]张鹏锐,王荣华.日本新一代多功能卫星MTSAT.中国航天,2005,4:10-12
[24]徐建平.日本新一代静止气象卫星MTSAT资料的接收处理.气象,1999,25(3):26-29
[25]张培昌,杜秉玉,戴铁丕.雷达气象学.北京:气象出版社,2000.11
[26]张培昌.多普勒天气雷达探测原理简述.山东气象,2000,20(3):9-13
[27]王令.多普勒雷达及其应用.江西气象科技,1999,22(1):28-32
[28]李腹广,王芬.用天气雷达回波强度资料估测降水.气象科技,2007,35(2):286-288
[29]王改利,刘黎平,阮征.多普勒雷达资料在暴雨临近预报中的应用.应用气象学报,2007,18(3):388-395
[30] Mecikalski J R,K M Bedka,S J Paech,and L A Litten.A statistical evaluation of GOES cloud-top properties for nowcasting convective initiation.Mon.Wea. Rev.,2008,136:4899-4914
[31]黄建明,李金宗,李冬冬.利用二位插值核的高分辨力图像重构.光电工程,2005,5:80-84
[32]陈宝平,赵俊岚,尹志凌.双线性插值算法的一种快速实现方式.北京电子科技学院学报,2004,12(4):21-23
[33]段汕,秦前清.基于形态插值的图像重建方法.计算机工程及应用,2004,18:26-28
[34]杨云峰,苏志勋,胡金燕.一种保持边缘特征的图像插值方法.中国图像图形学报,2005,10:1249-1251
[35] Ranchin T,Wald L.The Wavelet Transform for the Analysis of Remotely Sensed Images.Int. J. Remote Sensing,1993,14(3):615-620
[36] Li H,Manjunath B S,Mitra S K.Multisensor Image Fusion Using the Wavelet Transform. Graphical models and Image Processing.1995,7:235-245
[37]杨猛,胡卫东,夏胜平,郁文贤.一种基于规则的模式分类器设计方法.计算机工程与应用,2004,19:97-101
[38]李盼池,肖红,许少华,刘显德.支持向量机在模式分类中的应用.大庆石油学院学报,2003,6:59-61
[39] Browning K A,C G Collier.An integrated radar-satellite nowcasting system in the UK.Academic Press,1982:47-61
[40] Byers H R,R R Braham Jr.The Thunderstorm.U.S.Govt. Printing Office,1949:287
[41] Szejwach G.Determination of semi-transparent cirrus cloud temperature from infrared radiances.Application to METEOSAT. J. Appl. Meteor.,1982(21):384-393
[42] McCleese D J and Willon L S . Cloud top Height from temperature sounding instruments.Quant. J. R. Met. Soc.,1976(102):781-790
[43] Smith W L,C M R Platt.Intercomparison of radiosonde, ground based laser, and satellite deduced cloud heights.J. Appl. Meteor.,1978(17):1796-1802
[44] Nakajima T,Michael D King.Determination of the optical thickness and effective particle radius of cloud from reflected solar radiation measurements.Part I theory. J. Atmos. Sci.,1990(47):1878-1893
[45] Hansen J E,J B Pollack.Near-infrared light scattering by terrestrial clouds.J. Atmos. Sci.,1970(27):265-281
[46] Inoue T.An instantaneous delineation of convective rainfall area using split window data of NOAA-7 AVHRR.J. Meteor. Soc. Japan,1987,65:469-481
[47]李智勇,沈振康等.动态图像分析.北京:国防工业出版社,1999
[48]韩雷,俞小鼎,郑永光等.京津及邻近地区暖季强对流风暴的气候分布特征.科学通报,2009,54(11):1585-1590
[49] Wilks D S.Statistical methods in the atmospheric sciences.2nd ed. Academic Press,2006:627
[50] Bermowitz R T , E A Zurndorfer . Automated guidance for predicting quantitative precipitation.Mon. Wea. Rev.,1979,107:122-128
[51] Pearson K.On lines and planes of closest fit to systems of points in space.Philos Mag,1901,6(2):559-572
[52]施能.气象科研与预报中的多元分析方法(第二版).北京:气象出版社,2002:72
[53] Jolliffe I T.Principal component analysis.2nd ed. Springer,2002:487
[54] Preisendorfer R W . Principal component analysis in meteorology and oceanography.Elsevier,1988:425
[55] Pruppacher H R,J D Klett.Microphysics of Clouds and Precipitation.Kluwer Academic Publishers,1998:954
[2] Dixon M , Wiener G . TITAN: Thunderstorm identification, tracking, analysis, and nowcasting a radar-based methodology.Journal of Atmospheric and Oceanic Technology,1993,10(6):785-797
[3] Johnson J T,Mackeen P L,Witt A, et al.The Storm Cell Identification and Tracking algorithm: an enhanced WSR-88D algorithm.Weather and Forecasting,1998,13:263-276
[4] Jan Handwerker.Cell tracking with TRACE3D—a new algorithm.Atmospheric Research 2002,61:15-34
[5] K A布朗宁.现时预报(周凤仙,马振,李泽椿译).北京:气象出版社,1986,260
[6]许锐.基于卫星数据的对流初生自动识别研究:[硕士学位论文].青岛:中国海洋大学,2009
[7] Maddox R A.Mesoscale convective complexes.Bull Amer Meteor Soc,1980,61(11):1374-1387
[8] Setvák M,Doswell C A.The AVHRR channel 3 cloud top reflectivity of convective storms.Mon. Wea. Rev.,1991,119:841-847
[9] Strabala K I,Ackerman S A,Menzel W P,Cloud properties inferred from 8–12 _m data.J. Appl. Meteor.,1994,33:212-229
[10] Schmetz J,S A Tjemkes,M Gube,et al.Monitoring deep convection and convective overshooting with METEOSAT.Adv. Space Res.,1997,19:433-441
[11] Roberts R D,S Rutledge.Nowcasting storm Initiation and growth using GOES-8 and WSR-88D data.Wea. Forecasting,2003,18:562-584
[12] Hand W H . An object-oriented technique for nowcasting heavy showers and thunderstorms.Meteor. Appl.,1996,3:31-41
[13] Nair U S,R C Weger,K S Kuo,R M Welch.Clustering, randomness, and regularity in cloud fields 5. The nature of regular cumulus cloud fields.J. Geophys. Res.,1998,103(D10):11363-11380
[14] Purdom J F W,Menzel W P,Introducing GOES-I: The first of a new generation of geostationary operational environmental satellites.Bull. Amer. Meteor. Soc.,1994,75:757-781
[15] Beckman S K.Relationship between cloud bands in satellite imagery and severe weather. Satellite Imagery Interpretation for Forecasters.1986,Vol. 2,Precipitation Convection, P. S. Parke, Ed.,National Weather Association.[Available from the NWA, 4400 Stamp Road, No. 404, Temple Hills, MD 20748.]
[16]倪允琪,周秀骥,王德英等.我国重大天气灾害形成机理与预测理论研究进展.中国气象科学研究院年报,2002,1:8-12
[17]徐建平.国内外气象卫星发展.空间科学学报,2000,10(20):104-115
[18]陈渭民.卫星气象学(第二版).北京:气象出版社,2005,535
[19]李伟,张恒.气象卫星的应用与发展.卫星与网络,2007,8:56-60
[20]王敏.MTSAT卫星云图中对流初生的自动预报算法研究:[硕士学位论文].青岛:中国海洋大学,2010
[21]石秀云.卫星云图分析基础及应用.青海科技,2010,3:43-46
[22]徐建平.日本MTSAT多功能卫星.气象科技,2005,33(1):96-97
[23]张鹏锐,王荣华.日本新一代多功能卫星MTSAT.中国航天,2005,4:10-12
[24]徐建平.日本新一代静止气象卫星MTSAT资料的接收处理.气象,1999,25(3):26-29
[25]张培昌,杜秉玉,戴铁丕.雷达气象学.北京:气象出版社,2000.11
[26]张培昌.多普勒天气雷达探测原理简述.山东气象,2000,20(3):9-13
[27]王令.多普勒雷达及其应用.江西气象科技,1999,22(1):28-32
[28]李腹广,王芬.用天气雷达回波强度资料估测降水.气象科技,2007,35(2):286-288
[29]王改利,刘黎平,阮征.多普勒雷达资料在暴雨临近预报中的应用.应用气象学报,2007,18(3):388-395
[30] Mecikalski J R,K M Bedka,S J Paech,and L A Litten.A statistical evaluation of GOES cloud-top properties for nowcasting convective initiation.Mon.Wea. Rev.,2008,136:4899-4914
[31]黄建明,李金宗,李冬冬.利用二位插值核的高分辨力图像重构.光电工程,2005,5:80-84
[32]陈宝平,赵俊岚,尹志凌.双线性插值算法的一种快速实现方式.北京电子科技学院学报,2004,12(4):21-23
[33]段汕,秦前清.基于形态插值的图像重建方法.计算机工程及应用,2004,18:26-28
[34]杨云峰,苏志勋,胡金燕.一种保持边缘特征的图像插值方法.中国图像图形学报,2005,10:1249-1251
[35] Ranchin T,Wald L.The Wavelet Transform for the Analysis of Remotely Sensed Images.Int. J. Remote Sensing,1993,14(3):615-620
[36] Li H,Manjunath B S,Mitra S K.Multisensor Image Fusion Using the Wavelet Transform. Graphical models and Image Processing.1995,7:235-245
[37]杨猛,胡卫东,夏胜平,郁文贤.一种基于规则的模式分类器设计方法.计算机工程与应用,2004,19:97-101
[38]李盼池,肖红,许少华,刘显德.支持向量机在模式分类中的应用.大庆石油学院学报,2003,6:59-61
[39] Browning K A,C G Collier.An integrated radar-satellite nowcasting system in the UK.Academic Press,1982:47-61
[40] Byers H R,R R Braham Jr.The Thunderstorm.U.S.Govt. Printing Office,1949:287
[41] Szejwach G.Determination of semi-transparent cirrus cloud temperature from infrared radiances.Application to METEOSAT. J. Appl. Meteor.,1982(21):384-393
[42] McCleese D J and Willon L S . Cloud top Height from temperature sounding instruments.Quant. J. R. Met. Soc.,1976(102):781-790
[43] Smith W L,C M R Platt.Intercomparison of radiosonde, ground based laser, and satellite deduced cloud heights.J. Appl. Meteor.,1978(17):1796-1802
[44] Nakajima T,Michael D King.Determination of the optical thickness and effective particle radius of cloud from reflected solar radiation measurements.Part I theory. J. Atmos. Sci.,1990(47):1878-1893
[45] Hansen J E,J B Pollack.Near-infrared light scattering by terrestrial clouds.J. Atmos. Sci.,1970(27):265-281
[46] Inoue T.An instantaneous delineation of convective rainfall area using split window data of NOAA-7 AVHRR.J. Meteor. Soc. Japan,1987,65:469-481
[47]李智勇,沈振康等.动态图像分析.北京:国防工业出版社,1999
[48]韩雷,俞小鼎,郑永光等.京津及邻近地区暖季强对流风暴的气候分布特征.科学通报,2009,54(11):1585-1590
[49] Wilks D S.Statistical methods in the atmospheric sciences.2nd ed. Academic Press,2006:627
[50] Bermowitz R T , E A Zurndorfer . Automated guidance for predicting quantitative precipitation.Mon. Wea. Rev.,1979,107:122-128
[51] Pearson K.On lines and planes of closest fit to systems of points in space.Philos Mag,1901,6(2):559-572
[52]施能.气象科研与预报中的多元分析方法(第二版).北京:气象出版社,2002:72
[53] Jolliffe I T.Principal component analysis.2nd ed. Springer,2002:487
[54] Preisendorfer R W . Principal component analysis in meteorology and oceanography.Elsevier,1988:425
[55] Pruppacher H R,J D Klett.Microphysics of Clouds and Precipitation.Kluwer Academic Publishers,1998:954